OpenAI official logo (public domain, Wikimedia Commons) — CryptoBriefing brand treatment
OpenAI rolls out composer predictions in Codex for eligible Pro users
The beta feature costs nothing extra for now, and it adds to Codex's shift from narrow coding agent to broader development tool
OpenAI is adding another feature to Codex, its software engineering agent. Composer predictions in Codex are now in beta for eligible ChatGPT Pro users.
The price tag during the beta is zero, beyond what those subscribers already pay.
What OpenAI announced
The announcement, posted on X, was short. Composer predictions in Codex are available to eligible Pro users during the beta, at no extra cost.
The first qualifier is “eligible.” Not every Pro subscriber is necessarily included. OpenAI has not set out publicly which accounts qualify.
The second is “beta.” The feature is still being tested in the wild, so it may change, expand or shrink before any general release.
The third is “during the beta.” The no-extra-cost terms are tied to the test period. OpenAI has not said what pricing will look like once the beta ends.
OpenAI has not published a technical breakdown of how predictions are produced or how fast they arrive.
AI, tech, and the markets they move—in one daily briefing.
Daily. Free. Join 34,000+ readers across crypto, finance, and policy.
How Codex got here
Codex launched in May 2025 as a cloud-based software engineering agent, initially limited to select users and designed to execute software-engineering tasks in preloaded cloud environments.
Since then, OpenAI has expanded its functionality across various platforms, integrating its capabilities into command-line interfaces and IDEs, positioning Codex as a broader cross-platform development tool.
Codex Cloud is available to Plus, Pro, Business, Enterprise, Healthcare and Education accounts.
Composer predictions are not available across that whole list yet. For now, the beta is reserved for eligible Pro users, OpenAI’s top consumer tier.
What this means for developers and OpenAI
For Pro subscribers, the near-term calculation is easy. They get a new Codex feature to try without paying more.
Pricing is the key thing to track. The “no extra cost” language is explicitly tied to the beta. When the test period ends, OpenAI could keep the feature bundled into Pro, move it to other tiers, or attach a separate charge.
If composer predictions eventually reach Plus, Business, Enterprise, Healthcare and Education accounts — the same list that already has Codex Cloud — the feature would move from a premium experiment to a standard part of the product.